滚动轴承的寿命预测模型建立方法及寿命预测方法
By introducing modal parameters and health indices into the rolling bearing life prediction model, and combining sequence partitioning and regression learning, the problem of prediction difficulties under changing rolling bearing data distribution is solved, achieving accurate life prediction and enhanced generalization under changing scenarios.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUAZHONG UNIV OF SCI & TECH
- Filing Date
- 2023-01-13
- Publication Date
- 2026-07-17
AI Technical Summary
Existing rolling bearing life prediction methods struggle to accurately predict remaining life when the distribution of their own data changes. Traditional methods rely on a large number of full life cycle samples and different operating conditions, and traditional feature extraction cannot effectively perceive the degradation information of rolling bearings.
A life prediction model for rolling bearings is established. The time series data is divided into dissimilar time segments by a sequence segmentation module. Loading data is constructed using modal parameters and health indices. Regression learning is performed by combining recurrent neural networks to achieve self-transfer and time invariance, reduce distribution differences, and improve generalization.
When the data distribution of rolling bearings changes, the remaining life can be accurately predicted, reducing the requirements for samples and operating conditions, avoiding prediction drift, and improving the model's generalization performance and practical application value.
Smart Images
Figure CN116384540B_ABST